Profiling Tools

This directory contains tools for profiling and monitoring Redpanda Connect performance using Prometheus, Grafana, and pprof.

Quick Start

Start the monitoring stack:
bash
```
task up
```
Run your Redpanda Connect instance with the desired configuration.
Access the dashboards:
- Grafana: http://localhost:3000
- Prometheus: http://localhost:9090

Capturing Profiles

In order to use profiling make sure your Redpanda Connect instance has the following configuration:

yaml

http:
  debug_endpoints: true

Use the following Taskfile commands to capture different types of profiles:

bash

# Capture all profiles (CPU, memory, blocking)
task profile

# Or capture specific profiles:
task profile:cpu    # 30s CPU profile
task profile:mem    # Memory profile
task profile:block  # Goroutine blocking profile

Profiles are saved to the ./profiles directory, you can use the pprof tasks to open them in a browser:

bash

task pprof:cpu
task pprof:mem
task pprof:block

Cleanup

To stop and remove all containers:

bash

task down